Skip to content

Actions: facebookresearch/xformers

win-build

Actions

Loading...
Loading

Show workflow options

Create status badge

Loading
72 workflow runs
72 workflow runs

Filter by Event

Filter by Status

Filter by Branch

Filter by Actor

Integrate int4kv to benchmark_attn_decode.py
win-build #290: Pull request #1029 synchronize by scxiao
April 26, 2024 20:56 50m 44s scxiao:int4kv_fa_decode
April 26, 2024 20:56 50m 44s
Integrate int4kv to benchmark_attn_decode.py
win-build #289: Pull request #1029 synchronize by scxiao
April 26, 2024 19:57 52m 35s scxiao:int4kv_fa_decode
April 26, 2024 19:57 52m 35s
Integrate int4kv to benchmark_attn_decode.py
win-build #288: Pull request #1029 synchronize by scxiao
April 25, 2024 21:57 48m 41s scxiao:int4kv_fa_decode
April 25, 2024 21:57 48m 41s
Fast Multi-ahead Attention support on AMD ROCM
win-build #281: Pull request #978 synchronize by tenpercent
March 1, 2024 19:00 17m 37s ROCm:dev_upstream
March 1, 2024 19:00 17m 37s
Fast Multi-ahead Attention support on AMD ROCM
win-build #279: Pull request #978 synchronize by qianfengz
February 29, 2024 01:20 19m 56s ROCm:dev_upstream
February 29, 2024 01:20 19m 56s
Fast Multi-ahead Attention support on AMD ROCM
win-build #277: Pull request #978 synchronize by tenpercent
February 28, 2024 20:31 19m 27s ROCm:dev_upstream
February 28, 2024 20:31 19m 27s
Fast Multi-ahead Attention support on AMD ROCM
win-build #275: Pull request #978 synchronize by tenpercent
February 26, 2024 19:19 22m 10s ROCm:dev_upstream
February 26, 2024 19:19 22m 10s
Fast Multi-ahead Attention support on AMD ROCM
win-build #273: Pull request #978 synchronize by tenpercent
February 22, 2024 18:59 37m 16s ROCm:dev_upstream
February 22, 2024 18:59 37m 16s
Fast Multi-ahead Attention support on AMD ROCM
win-build #269: Pull request #978 synchronize by tenpercent
February 19, 2024 18:43 57m 13s ROCm:dev_upstream
February 19, 2024 18:43 57m 13s
Fast Multi-ahead Attention support on AMD ROCM
win-build #259: Pull request #978 synchronize by tenpercent
February 16, 2024 01:11 1h 3m 22s ROCm:dev_upstream
February 16, 2024 01:11 1h 3m 22s
Fast Multi-ahead Attention support on AMD ROCM
win-build #238: Pull request #978 synchronize by qianfengz
February 8, 2024 13:53 2m 31s ROCm:dev_upstream
February 8, 2024 13:53 2m 31s
Fast Multi-ahead Attention support on AMD ROCM
win-build #236: Pull request #977 opened by qianfengz
February 7, 2024 14:08 2m 34s qianfengz:dev_to_upstream
February 7, 2024 14:08 2m 34s
Fix bug in mem_eff kernel with attention mask and MQA
win-build #235: Pull request #952 opened by drisspg
December 21, 2023 16:21 57m 52s fix_mem_eff_masking_mqa_bug
December 21, 2023 16:21 57m 52s
Decouple the kernel_forward.h/kernel_backward.h from PyTorch Dependencies
win-build #228: Pull request #916 synchronize by Li-Jiajie
November 6, 2023 11:57 49m 40s Li-Jiajie:dropout
November 6, 2023 11:57 49m 40s
Decouple the kernel_forward.h/kernel_backward.h from PyTorch Dependencies
win-build #227: Pull request #916 synchronize by Li-Jiajie
November 6, 2023 11:44 58m 17s Li-Jiajie:dropout
November 6, 2023 11:44 58m 17s
Decouple the kernel_forward.h/kernel_backward.h from PyTorch Dependencies
win-build #226: Pull request #916 synchronize by Li-Jiajie
November 6, 2023 11:29 54m 51s Li-Jiajie:dropout
November 6, 2023 11:29 54m 51s
Decouple the kernel_forward.h/kernel_backward.h from PyTorch Dependencies
win-build #225: Pull request #916 synchronize by Li-Jiajie
November 6, 2023 11:25 58m 20s Li-Jiajie:dropout
November 6, 2023 11:25 58m 20s
Decouple the kernel_forward.h/kernel_backward.h from PyTorch Dependencies
win-build #224: Pull request #916 opened by Li-Jiajie
November 6, 2023 11:22 58m 13s Li-Jiajie:dropout
November 6, 2023 11:22 58m 13s